phi-3-orpo-v9_4

Uploaded model

• Developed by: cstr
• License: apache-2.0
• Finetuned from model : cstr/phi-3-orpo-v8_16

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>

EU AI Act Art. 53 — provider obligations

Added 2026-08-02 during an account-wide provenance review.

This is a fine-tune, not a format conversion. Most cstr/* repositories are GGUF conversions where the upstream research team remains the provider of the model and conversion changes only the numeric representation of the weights. Training changes the model itself, so under Regulation (EU) 2024/1689 the maintainer of this repository is plausibly the provider of the model it produced, and the obligations that survive the Art. 53(2) free-and-open-source exemption — Art. 53(1)(c) and 53(1)(d) — attach here.

Art. 53(1)(c) — copyright policy. This model shares its training pipeline with cstr/phi-3-orpo-v9_16, which used johannhartmann/mistralorpo. The specific run parameters for this variant are not recorded in this repository. No text or data mining was performed to assemble a corpus for this repository, so no rights reservation under Art. 4(3) of Directive (EU) 2019/790 was engaged by this step. Any credible claim that this model was trained on material without the necessary rights will be acted on — contact via the Community tab.

Art. 53(1)(d) — training content. ORPO steps on German preference data on top of cstr/phi-3-orpo-v8_16. This card records fewer run details than its sibling cstr/phi-3-orpo-v9_16, and the difference between the two variants is not documented here — stated as a gap rather than guessed at. The underlying training content is Microsoft's Phi-3-mini, documented by its provider.

Base model. cstr/phi-3-orpo-v8_16 — the training built on those weights and inherits their terms.